Nadir photography issue with verticals

[sparrowhawk360]

Hi guys

I've read a lot about shooting the nadir image including threads on this forum. However, they all seem to assume the same thing - a flat floor surface with no close rising vertical elements. IE chair backs etc.

Just to give a little context, I have a

MicroFour Thirds camera, Nodal Ninja Mk III, tripod and Panasonic Lumix 8mm fisheye (https://www.dpreview.com/reviews/panasonic-8-3p5-o20/3)

BTW this fisheye takes fullframe fisheye photos so there's only a 180 degree diagonal view. No neat little circular 180 x 180 image.

The end result of shooting panoramas like this is I end up with a 3-4 foot black hole at the nadir.

Where I have a completely flat surface, I've usually pointed the camera down about 45 degrees and pulled the tripod back about 3/4 feet. I get an angled downward shot which PTGui handles very well with it's "viewpoint correction".

Sometimes I mistakenly get a foot or tripod leg in a shot or dont drag the tripod back far enough but they're fixable with some photoshopping and masking

So the above method is one I read about most on these forums and works pretty well.

However, I was on a shoot recently where the angled nadir shot didn't work at all due to the nadit image containing the backs of a number of chairs and some steps.

Ideally this shot should have been taken pointing down at -90 degrees at the exact original spot as the other images. Unfortunatlely, that wasn't possible.

Any tips for this scenario? I have another shoot coming up soon of a conference hall with very close seating etc and need to practice.

Many thanks

[nick fan]

Well, the only way in your case is to reduce the nadir footprint and move away from 3D objects. You can tilt the lens down (eg 10-15 deg) for the horizontal shots to reduce the nadir hole. What is the size of your nadir footprint in terms of % width of your pano? What accounts most for the footprint? Tripod or rotator?

Nick

[sparrowhawk360]

Well, the only way in your case is to reduce the nadir footprint and move away from 3D objects. You can tilt the lens down (eg 10-15 deg) for the horizontal shots to reduce the nadir hole.
Nick

Will definitely try that. I suppose I'd end up with a larger zenith hole and might require a subsequent row photographed at 10-15 degrees up.

What accounts most for the footprint? Tripod or rotator?
Nick

The rotator I suppose.
I say "suppose" as I don't know - the nadir hole is always larger than tripod and rotator so they're never photographed.
I'd say when converted to cubes that the nadir hole is 50% of the cube.

[nick fan]

zenith is easy to deal with as it is at NPP.

Nick

[John Houghton]

The safest way to cope with objects such as chairs close to the tripod is to take two shots with the camera pointing straight down, with the pano head rotated by 180 degrees between them. This captures everything down below from the same viewpoint as all the other shots. This includes the tripod, of course, so you need an additional nadir shot of the small (flat) area the tripod occupies. You can shoot this from an offset position, as you have already done, though it's preferable to get the camera reasonably close to the original position either by leaning the tripod over with the back leg extended and your foot on it to prevent it toppling over, or a nadir adapter, or a horizontal pole. Matching floor shine can be a problem with offset shots and polished floors. But as Nick suggests, tilting the camera down for the horizontal shots can work well enough provided the image quality is good at the bottom of the image frame. With some lenses the image quality falls off significantly at the edge of the image circle and this leads to annoying abrupt changes in sharpness at the seams with the sharp centre of the nadir shot.

John

[Hugh]

Following on from John's good advice you may also like to consider revising your shooting pattern to minimise the hole at the nadir so that you are left with dealing with removing the head and tripod legs rather than having a large hole to fill.

I am making the assumption that you are tilting your rig upwards to close the zenith which is causing the "large" hole at the nadir.

The panoramas on http://360hugh.co.uk/panormas/seat-locations/ were made some years ago with a Nikon D300 and 10.5mm lens on an NN 3 with 8 shots round @ -10° + 4 shots round @ + 45°.
This also provided a lot of overlap for Control Point generation.

[sparrowhawk360]

Hi guys - thank you for all the advice.

I've tried my first tilted shots so instead of shooting 1 row or images (1 image at every 60 degree stop) I've tried 2 rows.
The first row is tilted up 30 degrees up. The second row is tilted down 30 degrees down.

Good news:
This did leave a much smaller nadir hole and also meant I didn't have to shoot a separate zenith.

Bad news
I was surprised to find that this titled method gave me some obvious stitching errors. Upon closer inspection it revealed that my NPP isn't sorted.
This is obvious because not only can I see errors but the nadir consisting of the Nodal Ninja rotator is clearly misaligned. See image



Clearly I need to sort this out and clearly shooting at a tilt is less forgiving than a single row panorama (please confirm).

It's all good though - I enjoy learning and perfecting my setup.

[Hugh]

It looks as if your lower rail setting is too far to the left when viewed from behind the camera.

I set up my lower rail by setting the upper rail vertical (-90°, looking vertically down) so it lines up with the standard so that you are looking at the Nodal Ninja logo on the NN3 through your viewfinder then you can move the camera and lens along the rail until the centre of the logo is in the centre of the image.
To fine tune you can take a shot then zoom in to ensure the logo is in fact in the centre of the image.

Your Panasonic Lumix 8mm fisheye is somewhat smaller than my Samyang 8mm but you may find that the diagrams on http://hugha.co.uk/NodalPoint/Index.htm#Samyang 8mm helpful in getting a good starting point for the NPP although I suspect that your upper rail setting is OK on your rig and that it is the lateral movement along the lower rail that is causing the difficulties.

[rjrose]

Sorry this is a few months late, but let me add something. I am sure this may set off a few people, and go against the received wisdom. Whatever.

I also shoot a Full-Frame fisheye; a 15mm on a Nikon D810. I take three rows, 60° stops, 6 frames around, plus 2 more ± 60° rows. This covers the zenith and nadir. PTGi has no trouble stitching the 18 images into a perfect equirectangular that rotates in the Flickr view. I don't take a separate nadir shot; the third row takes care of that. See this album: https://flic.kr/s/aHskE5cHEG

Now for the part that will really get people mad. I think it is perfectly fine to eyeball the NPP (it is probably near the aperture ring) and the center of the camera (look at the lens straight on) over the pivot point. Close is good enough. I don't think the picture you posted indicates a problem with how you set up the camera on the head. I think it is just some misplaced control points. If you are using a program that lets you manually set control points try setting them yourself and see if that solves the issue. In any event if you are going to patch over the nadir with something like a mirror ball, then who cares what your nadir looks like?

[John Houghton]

I also shoot a Full-Frame fisheye; a 15mm on a Nikon D810. I take three rows, 60° stops, 6 frames around, plus 2 more ± 60° rows. This covers the zenith and nadir. PTGi has no trouble stitching the 18 images into a perfect equirectangular that rotates in the Flickr view.

Well, not always perfect stitching. I looked at the first of your panoramas that had something reasonably close (a chair) and saw an obvious stitching error (image crop attached). This is exactly the sort of glitch that you would expect from using a head not set up as accurately as it might be. Also, most people will not find it acceptable to have the tripod and its shadow, and also your shadow, visible in panoramas. The head can be set up quite accurately in ten minutes or so. It only needs to be done once, so it's worth doing. With a head thus set up, you need only six shots round + one zenith + (2 nadir shots at 90 or 180 - optional) + nadir shot with the tripod shifted away. This basic shooting configuration can be varied according to circumstances but should be found generally satisfactory.

John